Avoca Hits $1B Valuation — AI Voice Agents Are Coming for America's Service Economy

Picture a rooftop in Phoenix, Arizona. It's late April, but the desert doesn't care about calendars — it's already pushing 105 degrees. Mike runs a twelve-person HVAC shop. He's up on that roof, hands greasy, swapping out a compressor unit. His phone buzzes in his back pocket. He can't answer it. He's holding a wrench in one hand and bracing himself against a duct with the other. The call goes to voicemail. Nobody leaves a voicemail anymore. That was a homeowner whose entire AC system just died, ready to write a check for $3,000 to whoever picks up first. She calls the next company on Google. Mike never hears from her.

This happens three, four times a day. Not just to Mike — to hundreds of thousands of service contractors across America. HVAC techs, plumbers, roofers, electricians. They're small operators, five to fifty employees, and the owner is usually the one with the wrench. There's no receptionist. There's no call center. There's a cell phone in somebody's pocket, and half the time it goes unanswered because the person who owns it is elbow-deep in ductwork.

The US home services market is worth $600 billion. That's not a typo. Six hundred billion dollars, most of it flowing through businesses that manage their schedules on whiteboards and track customer records in paper binders. These companies don't have IT teams. They don't have CRMs. Their tech stack is a phone, a clipboard, and maybe Google Calendar if they're feeling adventurous.

Into this market, an AI walked in. And someone just put a billion dollars behind it.

---

Chen didn't have a eureka moment. Avoca's co-founder came to this business the hard way — by spending time in the field, watching how service contractors actually operate, and seeing exactly how broken the basics were.

The problem was embarrassingly simple. These businesses couldn't answer their phones.

Avoca's AI agent handles the entire service workflow: call intake, scheduling, technician assignment, and follow-up.

After hours, on weekends, on holidays — roughly 60% of incoming calls to US service businesses go unanswered. And here's the thing about a homeowner with a broken AC in Phoenix in summer: they're not patient. They call the first company, nobody picks up, they call the second. Whoever answers first gets the job. The company that missed the call doesn't just lose that job — they lose that customer forever. The homeowner isn't going to circle back.

So Chen tried the obvious thing first, right? Give them better software. A CRM. A scheduling tool. Something to organize the chaos. But these owners don't adopt software. They won't sit through an onboarding call. They won't watch a tutorial. ServiceTitan — which IPO'd in 2024 at a $9.5 billion market cap — built a comprehensive SaaS platform for this exact market, and it's a great product. But it costs thousands per month and has a learning curve. The big contractors use it. The long tail of smaller operators? They can't afford it, or they just won't bother.

Chen went a different direction. Don't sell them software. Answer their phone.

That's it. That's the whole pitch. "Forward your business number to us. The AI picks up." No installation. No login. No dashboard. No onboarding. Just phone number forwarding — something any contractor can set up in two minutes. From that moment, the AI answers every call. It listens to the problem, schedules the appointment, assigns a technician, quotes the job.

And that idea is now worth a billion dollars.

In April 2026, Avoca closed its Series B. Meritech Capital and General Catalyst co-led the round. Kleiner Perkins, which backed the Series A, followed on. Cumulative funding crossed $125 million. Post-money valuation: $1 billion. Officially a unicorn.

Metric	Value
Valuation	$1B (post-money)
Cumulative funding	$125M+
Series B leads	Meritech Capital + General Catalyst
Customers	800+ US service contractors
ARR	Eight figures (as of Dec 2025)
2026 job routing target	$1B
Target verticals	HVAC, plumbing, roofing, electrical

The company crossed eight-figure ARR in December 2025, and it's set a target to route $1 billion in jobs through its platform in 2026. One billion dollars in work, dispatched by an AI that started by answering the phone.

---

But hold on. A phone-answering AI is worth a billion dollars? That sounds... a lot, doesn't it?

To understand why the money makes sense, you need to understand the market — and more specifically, why every previous attempt to modernize it fell short.

Think about the typical HVAC contractor. Mike's shop has twelve people. Mike is the owner, the lead technician, and the de facto sales team. His wife handles the books. His most experienced tech, Danny, runs a second crew. There are ten other guys. Nobody in this operation knows what an API is. Nobody has ever heard of Salesforce. The scheduling system is a whiteboard in the break room with magnets that represent each tech's availability.

Now imagine walking into Mike's office and saying, "Hey Mike, I've got this incredible SaaS platform. It's got CRM, automated scheduling, dispatching, invoicing, marketing automation..." Mike's already tuned out. He's thinking about the compressor he needs to replace at the Johnson house tomorrow morning. Software is something that happens to other people.

But imagine walking in and saying, "Mike, forward your phone to this number. That's it. We'll answer every call you miss." Mike can do that. Mike does that. And from that day forward, Avoca's AI picks up every call to Mike's business. "Hi, thanks for calling Mike's HVAC. How can I help you today?" It listens to the homeowner describe the problem, checks the schedule, finds an open slot, assigns the right technician, and confirms the appointment. When Mike climbs down from the roof, there's a new job in the queue. The $3,000 that used to evaporate into a missed call is now revenue.

Here's where it gets clever. The phone call isn't really the product. It's the Trojan horse. Because every call Avoca handles generates data — customer records, job histories, scheduling patterns, technician performance metrics. Over weeks and months, Avoca accumulates a complete operational picture of Mike's business. It knows which tech is fastest at duct replacements. It knows which zip codes generate the most emergency calls on weekends. It knows that Thursday afternoons are slow and could be backfilled with maintenance appointments.

Chen put it this way in the company's blog post: "We're not building a chatbot. We're building the operating system for service businesses."

That's not marketing fluff. It's literally what's happening. Call answering leads to scheduling. Scheduling leads to dispatching. Dispatching leads to quoting. Quoting leads to invoicing. Each link in that chain is a new product surface, and each one deepens the data moat. By the time a competitor shows up, Avoca has six months of Mike's operational history — and migrating that is painful enough that Mike won't bother.

---

This is where Mamoon Hamid comes in.

Mamoon is a General Partner at Kleiner Perkins — one of the oldest and most storied venture firms in Silicon Valley. They backed Google. They backed Amazon. They backed Netscape. Mamoon led Avoca's Series A and followed on in this Series B.

In a Fortune interview, he said something that cuts through a lot of the AI hype: "Avoca is the rare vertical AI company that found product-market fit before raising big."

That single sentence is worth sitting with. The typical AI startup playbook in 2025 and 2026 goes like this: raise a massive round on a thesis ("AI will transform X!"), then go find customers to validate that thesis. Avoca flipped it. They had 800 paying contractors before the big money showed up. The customers came first. The valuation followed.

At eight-figure ARR and a $1 billion valuation, the revenue multiple is roughly 20-40x. That's high by traditional SaaS standards, but within bounds for a fast-growing AI-native company. The more interesting number is the $1 billion job routing target for 2026. That hints at a potential shift from pure subscription revenue to a take-rate model — skimming a percentage of every job that flows through the platform. Think of how Uber takes a cut of every ride. If Avoca moves toward transaction-based revenue, its TAM explodes beyond what monthly subscriptions could ever reach.

---

So where does Avoca sit in the landscape? There are other voice AI companies. There are other home services software companies. But Avoca occupies a peculiar position that neither group fully overlaps with.

Company	Position	Valuation/market cap	Differentiator
ServiceTitan	Home services SaaS platform	$9.5B (IPO 2024)	Full SaaS, targets larger contractors
Jobber	Field service management	$1B+	Generalist field service
Housecall Pro	Home services management	$600M+	ServiceTitan competitor
Bland AI	General voice AI	$60M+ funding	Not vertically specialized
Retell AI	Voice AI infrastructure	$30M+ funding	Developer tools
Vapi	Voice AI platform	$40M+ funding	API-first

Bland AI, Retell AI, and Vapi are horizontal voice AI infrastructure. They sell APIs to developers. "Build any voice agent you want." That's powerful if you're a technical buyer. But Mike in Phoenix doesn't have developers. He has a phone and a clipboard.

In the voice AI landscape, Avoca occupies a unique position: vertically specialized with full workflow coverage.

Hand Mike a Bland AI API key and tell him to design a prompt, integrate with his scheduling system, and set up a workflow — he'll stare at you like you just spoke Mandarin. He can fix a compressor in his sleep, but he doesn't know what a webhook is.

Avoca's pitch to Mike: "Forward your calls to us. We'll handle the rest." Done. No code. No setup. No learning curve. That's product-market fit for a non-technical buyer, and it's the reason Avoca's valuation is ten times larger than the horizontal players despite operating in a narrower market. Narrow beats wide when your customers can't customize the wide option.

This valuation gap tells a clear story. Bland AI has $60 million in funding. Avoca has a $1 billion valuation. They're both "voice AI companies," but the market is pricing vertical specialization at a massive premium over horizontal infrastructure. When your customers are small business owners who can't code, a ready-made solution beats a toolkit every time.

---

The relationship with ServiceTitan is worth unpacking.

ServiceTitan is the gorilla in this market. It IPO'd in 2024 at a $9.5 billion market cap. Full SaaS platform — CRM, scheduling, dispatching, invoicing, marketing, the works. It's the dominant player for larger contractors.

But Avoca and ServiceTitan aren't really competing. Not yet.

ServiceTitan's sweet spot is contractors with 50-plus employees, millions in annual revenue, and a budget for enterprise software. These are operations that have office managers, dispatch coordinators, maybe even a marketing person. They can handle a complex SaaS platform.

Avoca is going after the tier below. Five to thirty employees. Hundreds of thousands in revenue. Minimal software adoption. These companies find ServiceTitan too expensive or too complex. By company count, this segment is vastly larger — there are far more small contractors than big ones.

So for now, it's complementary. Avoca picks up the customers ServiceTitan can't reach. But the long game is different. As Avoca expands from call answering into full workflow management — scheduling, dispatching, quoting, invoicing — it inevitably moves upmarket into ServiceTitan's territory. And there's a structural advantage at play: an architecture built from scratch around AI versus a legacy SaaS product with AI bolted on after the fact. That gap widens over time.

If you're a ServiceTitan shareholder, start tracking Avoca's quarterly customer count. The upmarket migration has already begun.

---

Zoom out further and Avoca starts to look like something bigger than a home services play. It's defining a new category: AI agents for SMBs.

The AI agent conversation has been dominated by enterprise use cases. Microsoft Copilot, Salesforce Einstein, Google Duet — tools for big companies with complex workflows and knowledge workers. But 99% of the US economy is small businesses. Thirty-three million of them. Restaurants, dry cleaners, auto repair shops, dental offices, nail salons, law firms. These businesses don't use Notion. They don't use Slack. Their primary tools are a phone and a paper calendar.

What does an AI agent look like for this world? Avoca answered that question.

First, the entry point has to be the phone. SMB owners won't install new software. But they'll forward a phone number. It's an extension of existing behavior, not new behavior. That's a crucial distinction — the best SMB products enter through something the customer is already doing, not something they need to learn.

Second, the AI fills gaps instead of replacing people. Most service businesses miss calls after hours, on weekends, and during holidays — not because they don't want to answer, but because nobody's there. Up to 60% of those calls never convert to revenue. Avoca doesn't replace the receptionist; it covers the hours the receptionist isn't working. This framing matters. "We'll fire your receptionist and replace her with AI" is a terrifying pitch. "We'll catch the calls nobody's catching" is a relief.

Third, workflow data compounds into a platform. Call answering generates scheduling data. Scheduling generates dispatch data. Dispatch generates performance data. Each layer makes the next layer smarter and stickier. After six months, Avoca knows Mike's business better than Mike does — which zip codes generate the most weekend emergencies, which technician closes the most upsells, which time slots go unfilled. That data becomes the foundation for entirely new services: insurance recommendations, parts procurement, financing for customers who can't pay upfront. A single phone call seeds a data flywheel that eventually powers an ecosystem.

---

The timing of this round intersects with other forces moving the same direction.

DeepSeek V4-Pro's 75% price cut is directly good news for companies like Avoca. Every phone call Avoca's AI handles costs inference money — the model is consuming tokens for the entire three-to-five-minute conversation. When input token costs drop 75%, gross margins improve immediately. That means Avoca can either lower its subscription price to be more aggressive with SMBs, or keep the price and pocket better margins. The cost compression at the model layer flows straight into vertical AI unit economics. It's invisible to the contractor, but it fundamentally changes the math.

The $600B US home services market has single-digit AI penetration. Avoca has room to run.

Bezos's Prometheus and its $38B round also connects. Prometheus is betting on physical AI foundation models — AI that can interact with the real world. Avoca starts with voice, which is software AI. But the endpoint of home services is physical: a technician on a roof, a plumber under a sink. If physical AI materializes — AR glasses showing repair guides, drones inspecting roofs before the tech climbs up, robots hauling parts — then vertical platforms like Avoca are the natural first integration point. They already have the workflow data. They already know which job is happening where, with what equipment, performed by which technician. The digital scheduling system meets the physical execution layer.

And then there's the macro tailwind. ServiceTitan's IPO in 2024 proved the home services tech category was real. A $9.5 billion public company validated that investors should care about plumbers and HVAC techs. On that foundation, Avoca emerged as the AI-native next generation. This is a pattern tech repeats every cycle. The first-generation SaaS company creates the market. The second-generation AI-native company reshapes it. Salesforce created CRM. AI-native CRMs are now coming for that market. ServiceTitan created home services SaaS. Avoca is coming for that market with an AI-native approach.

---

So who wins and who loses?

The Avoca team is winning. Chen, co-founder Shrivastava, and Braswell just hit unicorn status. Their early investors are winning. Mamoon Hamid at Kleiner Perkins watched his Series A thesis play out exactly as he predicted. The firm that backed Google and Amazon can now add "the company that AI-ified America's plumbers" to its portfolio story.

Small service business owners are winning too. For the first time, they have an AI tool that doesn't require them to become technologists. Mike couldn't afford a $40,000-a-year receptionist. Now he pays a few hundred dollars a month and every call gets answered, 24/7. That's a genuine improvement in his life and his revenue.

Horizontal voice AI startups got an uncomfortable signal. Bland AI, Retell AI, Vapi — look at their valuations. Tens of millions. Avoca just got a billion. Same category label, ten-to-one valuation gap. The market is pricing vertical specialization at a massive premium. These companies need to decide: stay horizontal and compete on infrastructure (where model cost drops from DeepSeek make it increasingly commoditized), or pivot vertical and compete on workflow depth. That decision point just arrived.

ServiceTitan is watching. The segments are different today, but Avoca is climbing upward. ServiceTitan's board is probably already debating: do we build our own AI voice agent, or do we acquire Avoca before it gets too expensive? That decision might define ServiceTitan's next five years. And if Google or Amazon decide to launch their own SMB AI agent platforms — which is entirely possible — the game changes for everyone.

Traditional call center outsourcing firms are losing. When an AI can answer the phone better than a human call center rep — and at a fraction of the cost — the value proposition of outsourced human call handling collapses. That's already happening.

---

Of course, there's a skeptic's case. And it's worth taking seriously.

Tomasz Tunguz — GP at Theory Ventures, former Redpoint partner, one of the sharpest SaaS analysts in the business — has pointed out the fundamental trap of SMB markets: churn. "The core problem with selling to small businesses is churn," he's said. "A contractor subscribes for $200 a month, tries it for 30 days, and cancels if the ROI isn't immediately obvious. AI voice agents have to prove they generate revenue — in dollars, not demos — or the 800 customer base plateaus before it becomes 8,000."

He's right, and this is probably the single biggest risk Avoca faces. SMB markets are easy to enter and easy to churn out of. Mike tries Avoca for a month. Did it actually generate revenue? If the answer is vague — "well, we answered some calls, but I'm not sure how many became actual jobs" — Mike cancels. He's not sentimental about SaaS subscriptions. For Avoca to beat this, they need to put hard numbers in front of every contractor every month: "You missed 47 calls last month before Avoca. We caught 38 of them. 12 converted to jobs. Total revenue captured: $14,200." If they can do that consistently, churn stays low. If they can't, Tunguz's warning becomes prophecy.

There's also a regulatory dimension. The FCC regulates AI-initiated calls to consumers. Avoca is currently inbound — it answers calls, it doesn't make them — so TCPA (Telephone Consumer Protection Act) constraints don't apply. But the moment Avoca starts making outbound calls — appointment reminders, follow-up surveys, promotional offers — state-level AI disclosure laws come into play. Several states now require AI callers to identify themselves as artificial, and the regulatory patchwork is growing. Navigating that while scaling nationally adds complexity and legal cost.

And voice AI technology itself isn't perfect yet. A Texas contractor with a thick drawl, a call with an AC unit roaring in the background, a furious homeowner talking fast — these are hard scenarios. If the AI mishears a detail and sends a tech to the wrong address, or classifies an urgent gas leak as a routine inquiry, the contractor's trust shatters instantly. The gap between 95% accuracy and 99% accuracy is where the hardest engineering problems live, and it's where customer trust is won or lost.

---

Still, looking at the big picture, what Avoca has built is compelling.

If you're building products for SMBs, the Avoca playbook has a core lesson: the entry point matters more than the feature set. SMBs don't buy software as software. They buy outcomes through channels they already use. For American contractors, that channel is the phone. For a Korean market, it might be KakaoTalk. For Southeast Asia, it could be WhatsApp. The principle is the same: enter through existing behavior, not new behavior. The "no-install entry point" is the single most important design decision for any SMB product.

If you're a voice AI startup, the valuation gap between horizontal and vertical should alarm you. DeepSeek and its competitors are driving inference costs toward zero, which means horizontal voice AI infrastructure is heading toward commodity pricing. "Our API is cheaper" isn't a moat. Vertical specialization with full workflow coverage is. Pick a vertical, go deep, own the customer relationship end-to-end.

If you're an investor, "SMB AI agents" is a category now. It has its first unicorn. Home services was the first domino. What's next? Legal — an AI that answers the phone at a law firm, triages inquiries, and schedules consultations with the right attorney. Dental — an AI that handles appointment booking, checks insurance eligibility, and sends pre-visit forms. Real estate — an AI that fields property inquiries, matches callers to listings, and schedules showings. Each vertical will produce its own Avoca-like winner. The first domino fell; the rest are a matter of time.

---

Let's go back to Phoenix. Back to the rooftop.

Mike uses Avoca now. When a call comes in while he's up on a roof, it doesn't matter. The AI picks up. "Thanks for calling Mike's HVAC, how can I help?" It listens to the homeowner. Checks the schedule. "We can have a technician out tomorrow morning at 10 — does that work for you?" When Mike climbs down, there's a new job in the queue. The $3,000 doesn't disappear anymore.

Mike probably doesn't think of it as "AI." He might not even know what a large language model is. He just knows that something answers his phone now and the missed calls stopped. That's exactly what Chen wanted. The best technology is invisible. It doesn't ask you to learn it. It doesn't ask you to change. It just works, quietly, in the background, catching the money you used to lose.

An AI answers the phone for plumbers now. A billion dollars says it won't stop there. And honestly? Looking at how many rooftops are left to climb, a billion might be just the beginning.

---

References

1. Avoca Blog — Series B Announcement
2. Fortune — Avoca Reaches $1B Valuation
3. PR Newswire — Avoca Series B Details